The 90-minute review

Weeks 1, 3, 4, 5, 6, 8, 9 and 10

Choose the genre for the contribution

• Research paper: support a new claim.
• Communication: present a focused, timely finding.
• Review: synthesize existing evidence.
• Data paper: enable dataset reuse.
• Technical report: inform a client or institution.
• Thesis: demonstrate research competence.

Read structure as evidence of purpose

• IMRaD: a question answered through methods and results.
• Thematic headings: evidence organized around an argument.
• Dataset description: enough context for reuse.
• Executive summary: a decision maker needs the main message first.

Use several structural clues together. Length alone does not establish the genre.

Fit the venue and the reader

Check four things before drafting:

• Audience: who needs this contribution?
• Scope: does the venue publish this problem?
• Article type: does it accept this form of contribution?
• Evidence: can readers inspect and reuse the work?

State one central claim and the evidence that warrants it.

Build a search from concepts

System AND process AND metric

Example concept blocks:

• Zinc anode OR Zn anode
• Dendrite growth OR dendrite suppression
• Cycle life OR coulombic efficiency

Use OR within a concept and AND between concepts. Add spelling variants, older names, and abbreviations. Check database syntax before running the query.

Keep a search log and repair metadata

Search log: database, date, full query, filters, hit count.

Reference record: verify authors, title, year, journal, and DOI against the source.

Capture → check → organize → cite.

Zotero stores and formats records. Accurate records and source verification make the bibliography dependable.

Screen first, then map the evidence

Set inclusion criteria before screening.

1. Title and abstract: include, exclude, or unsure.
2. Full text: resolve uncertainty and record exclusion reasons.

Build one row per study:

Claim · system · conditions · metric with units · limitation

Compare claims only after checking that conditions and definitions are comparable.

Compare the conditions before the claims

Pair task, 60 seconds: Do these outcomes establish a contradiction?

Lifetime gain from additive X versus control:

• Study A: 1 mA cm⁻², 1 mAh cm⁻² → +100 cycles.
• Study B: 10 mA cm⁻², 5 mAh cm⁻² → 0 cycles.

Synthetic teaching data. Uncertainty is unreported.

Worked answer: the regimes differ

These outcomes do not establish a contradiction.

• Current density and areal capacity both change.
• Uncertainty and failure definitions are missing.
• Keep each outcome within its tested conditions.

Whether the additive’s effect changes remains unresolved.

Next step: compare treatment and control while varying one condition systematically.

IMRaD gives each section a job

• Introduction: establish a worthwhile, unresolved question.
• Methods: make the answer checkable.
• Results: report the observations.
• Discussion: explain meaning and bound the claim.

Check alignment: aim → method → result → interpretation.

Every link needs a traceable connection to the next.

Reverse-outline paragraph functions

For each paragraph:

1. Number it.
2. Name its function in ten words or fewer.
3. Start with a verb.
4. Read the functions without the original text.

Topic: zinc dendrites.

Function: narrows the problem to high-current zinc operation.

Repeated functions can reveal a stalled argument.

Name the break and the repair

• Orphan aim: no matching result. Add evidence or narrow the aim.
• Unsupported claim: interpretation outruns results. Add evidence or limit it.
• Invisible method: a reported number lacks a defined procedure.
• Doubled move: paragraphs repeat a job.
• Buried gap: the unresolved problem arrives too late.

An Introduction moves toward one question

1. Establish a relevant context.
2. Narrow to the specific system.
3. Synthesize what is known and where it stops.
4. State the gap and its consequence.
5. State an aim that a result can answer.

Each move should advance the argument. Avoid narrowing to the same point twice.

Turn an absence into a consequential gap

Absence: Few studies report blended-solvent degradation under realistic flue gas.

Gap: Without degradation rates under those conditions, solvent replacement costs cannot be estimated reliably.

Build the argument in three steps:

Known → not known → consequence.

Identify the prediction, explanation, or decision that missing evidence prevents.

Write an aim that a measurement can settle

Weak: Understand solvent degradation.

Operational: Measure the degradation rate of two solvent blends at two oxygen levels under a fixed temperature and gas composition.

This aim requires:

• Defined blends and conditions.
• A degradation metric and calculation.
• Results for the stated comparisons.

Synthetic teaching example.

Methods must produce every reported quantity

Make the work checkable:

• Materials, equipment, and settings.
• Preparation and measurement sequence.
• Treatment, control, and matched conditions.
• Independent units, repeats, and exclusions.
• Metric definitions, calculations, and uncertainty.

Define failure, efficiency, and stable operationally. An instrument name cannot replace a complete procedure.

Identify independent experimental units

Pair task, 60 seconds

Three film batches are prepared independently per formulation. Four coupons are tested from each batch.

Treatment: additive X. Control: the same formulation without X.

For formulation comparison:

• What is the independent n per group?
• Which preparation and test conditions must be matched?

Synthetic teaching example.

Worked answer: batches are the independent units

n = 3 independent batches per formulation.

The 12 coupons are subsamples.

• Analyze batch summaries or account for nesting.
• Match polymer lot, casting, thickness, conditioning, and test settings.
• Define exclusions and report uncertainty.

Make the control comparison and the meaning of n explicit.

Results report the trade-off

At 2.0 s, raising temperature from 300 to 375 °C increased mean conversion by 40 percentage points and reduced mean selectivity by 25 percentage points.

These measurements show a trade-off within the tested range.

Identify independent experimental units

Repeated readings help assess measurement repeatability. They do not create independent experiments.

State the sampling unit, not just a number.

Practice: name the change precisely

Work alone, then compare with a partner.

Conversion changes from 40% to 80%. Selectivity changes from 90% to 65%.

For each response, write:

1. The change in percentage points.
2. The relative percentage change, using the initial value as the denominator.

Which wording makes the denominator explicit?

A caption defines every mark

Figure 1. Conversion of A and selectivity to B versus temperature at a residence time of 2.0 s. Points are means of three independent runs per temperature. Bars show ±1 sample SD. Lines guide the eye. Values are synthetic teaching data.

Check: quantity, conditions, summary, uncertainty, independent unit and other marks.

Start with one row per independent run

Keep conversion and selectivity paired within each run. Preserve the original table.

Check the grouped summary before plotting

Check one group by hand. Keep means and SDs in the same temperature order. Then label, export and reopen the figure.

Keep a package that recreates the figure

Keep together:

• Original run table and its units
• Analysis and figure instructions
• Required tools, versions and palette
• Caption and exported figure

Rebuild from a fresh start. Confirm that the same inputs recreate the summary and figure without hidden steps.

Another person should be able to follow the package.

Inspect the exported figure at its final size

• Prefer vector output for lines and text.
• For raster output, use 600 dpi or more at the intended size.
• Check labels, markers, bars and embedded fonts.
• Use symbols or direct labels to preserve meaning in grayscale.
• Use color consistently: teal for a reference; amber for the effect.

Conversion, selectivity and estimated yield

For this 1:1 reaction, estimated yield = mean conversion × mean selectivity / 100.

At 2.0 s, 350 °C has the highest estimate among those tested. The preferred condition depends on the objective and constraints.

Compare literature on the same basis

Before claiming agreement or improvement, check:

• Feed, catalyst, reactor and operating range
• Definitions and denominators for conversion, selectivity and yield
• Measurement basis, independent units and uncertainty
• Differences that could explain the comparison

Cite a verified source. If evidence is missing, state what must be found; do not invent a literature result.

A limitation must change the claim

A useful limitation changes interpretation or identifies the next test.

A conclusion answers within the tested scope

Use three moves:

1. Answer: state the central finding and magnitude.
2. Bound: name the conditions and remaining uncertainty.
3. Implication: state the decision supported, or the next useful test.

Keep every result traceable to Results. A Conclusion should not introduce a new result or an untested promise.

Practice: turn a claim into a conclusion

Revise this claim in two sentences:

375 °C is the best condition because conversion is highest. The selectivity trend proves that B decomposes.

Include:

• One numerical comparison and the tested condition
• A defensible implication or next test

Use the table: estimated yield is 54.6% at 350 °C, versus 52.0% at 375 °C.

Worked answer: evidence sets the boundary

At 2.0 s, the yield estimate from mean conversion and selectivity was highest among the four temperatures at 350 °C: 54.6%, versus 52.0% at 375 °C.

This comparison identifies a condition for further testing; a controlled residence-time study with byproduct measurements could test further reaction of B as one possible explanation.

Apply the review to one paper

• Trace aim → method → result → claim.
• Check the independent unit, n, uncertainty and denominator.
• Read each figure with its caption and original data.
• Compare studies on an explicit basis.
• Write one scoped finding and one next test with a predicted outcome.